Galapagos - save solution?

Replies to This Discussion

Permalink Reply by David Rutten on November 26, 2010 at 12:17am

Hi Wim,

no, not yet, sorry. Saving out Galapagos species to a file is on the list.

I don't know why you're getting a lot of NaNs, your network must be failing for certain slider positions. It might be a logical problem (divide by zero, overflow error), or it might be a bug in one of my components or it might be a bug in one of our Rhino SDK functions.

--
David Rutten
david@mcneel.com
Seattle, WA

Permalink Reply by Wim R.M. Dekeyser on November 26, 2010 at 1:25pm

Hi David,
OK, no problem, I'll be looking forward to you getting to that point on the wish list - we're getting close to Christmas anyway... :-)

About the NaN's. I didn't get any in the last run (397 generations of 50 individuals) of the same model that I got NaN's on before. As far as I can tell, it shouldn't be a division by zero problem, I'm solving the following formula: (a/10*b)-(c*d)-(e*f)-(g*h). I did save the text outcome of the run but I'm not sure which gene is what so I'm not sure how to recreate the exact situation... Is there something you want me to try?

BTW, what is a collision?

cheers,
wim

Genome[24], Fitness=-300.29, Genes [3% · 33% · 98% · 7%]
{
Record: Point Mutation at index 3: 0.0465 -> 0.0497
}
Genome[25], Fitness=NaN, Genes [3% · 33% · 98% · 9%]
{
Record: Genome was mutated to avoid collision

Permalink Reply by WT4 on August 11, 2016 at 2:20am

How do you get the parameters to appear in the record tab?? I cannot seem to find a way of doing this.  Many thanks

Permalink Reply by Richard Schaffranek on August 11, 2016 at 4:35am

Hallo Wim, 

you could use a datasaver with a limit of e.g. 50 this should somehow save you the last 50 genoms.

Or you can develop a custom c# component reading every solution and filtering according to some definition.

I will provide some code samples later today...

Permalink Reply by WT4 on August 11, 2016 at 5:34am

That would be great. Thank you. William